memoQ 6

At Kilgray, we maintain a single list of planned features. Some of these features are our own ideas, while others are ideas suggested by customers. I must confess that many customers became very clever in pointing out the shortcomings of existing functionality, so when it comes to improving functionality, that often comes from the customers. The rest comes from our 5-year roadmap which we have managed to maintain since 2004 (I mean... it's been updated :)). When we come together for deciding on what should go into the next version, we have two levels of discussions. The higher level is what we call concepts. These are large features, and previous examples include versioning, online projects with desktop documents, LiveDocs, or cascading filters. (Older concepts include the revamped term base for 3.0, views, folder structure management with export path rules, etc.) The lower level is what we call features. Some of the features can be larger (term extraction, subsegment leveraging, etc.), while others are very minor (for example selecting which term bases you write the term into, or a 'go to segment number' feature). We usually pick two or three concepts and a number of features that seems reasonable to deliver by the time we plan to release the next version. We have had the 6.0 meeting about two months ago, and then we put the list together, more or less.

The problem with this first meeting is always that we have more ideas than resources. So we draft a table with "Impact" and "Dev effort", and those that are large and large get priority, etc. Impact is usually more important than development effort, but we have to remain reasonable. We also attribute the usefulness of a feature or concept to user groups, and try to keep an even distribution (and of course prioritize things that benefit translators, language service providers and enterprises alike).

Today we had another meeting, a more technical one, which is about planning the ways the concepts should be implemented. This is very tiresome. I am writing this blog entry because my brain is squeezed out. It is the most intense thinking, reasoning, planning, and this is one of the reasons why I really enjoy my job. Building a business is nice, but includes a lot of routine tasks, but here ... there's everything but routine!

We decided that 6.0 will have three very large concepts (one of them is so large that we could hardly find a name for it -- so temporarily we called it project pimping which probably won't make it into the mainstream product :)), three large and important features, and a massive amount of small improvements. As both 4.5 and 5.0 were very major releases, we did not have the time to fix small items. The next version will also be a very major release, but now we made a pledge to address about 30 small items that relate mostly to the translator pro version.

We introduced a new category which we call "support helper", and that will be quite focal in the next version. There are a number of repeating support issues which are not the fault of memoQ but are related to certain Windows shortcomings. Obviously, we cannot call these memoQ bugs, so we don't fix them when they appear. However, there are two that are very frequent, the one that relates to functionality that is there in .NET Framework 2.0 SP2, but not there in the non-service packed version, and the other relates to our preferences.xml file that sometimes gets corrupted by antivirus software. Both of them are easy to fix - you only need to install the new .NET version in the first case, and you need to locate and delete preferences.xml. But these mean extra work for our support services, and also extra frustration for customers. Another thing that we want to improve on is the server installation time. In a simple environment, installing a server should not take more than 30 minutes or so, however, there are some prerequisites. We install 10 or more servers every week, and sometimes new customers come a bit unprepared. In cases like these, it takes up to 3-4 hours. This isn't good. So our objective is to give tools that halve the average installation time and reduce the risk of misconfiguration. These will come in 6.0.

What else? Well, I can't really tell details, but take it for granted that we'll continue to surprise people. On the freelance side, we continue our good tradition of interoperability, we improve the usability of the software, and we finally give you the long-awaited subsegment leverage. On the LSP and enterprise side? Please don't ask yet! :)